1

Arguments for

Atheism is the rejection of : a‐theism. Atheists maintain some or all of the following claims: that theism is false; that theism is unbelievable; that theism is rationally unacceptable; that theism is morally unacceptable.

Among arguments for atheism, there are arguments that are direct, indirect, and comparative.

Direct arguments for atheism aim to show that theism fails on its own terms: theism is meaningless, or incoherent, or internally inconsistent, or impossible, or inconsistent with known fact, or improbable given known fact, or less likely than not given known fact, or morally repugnant, and so forth.

Indirect arguments for atheism depend upon direct arguments for something else. Consider . Naturalism and theism are jointly inconsistent: they cannot both be true. Direct arguments for naturalism—arguments for the claim that naturalism is true, or rationally required, or morally required—are, eo ipso, arguments for atheism.

Comparative arguments for atheism are arguments for the theoretical superiority of something else to theism. Consider naturalism. An argument for the theoretical superiority of naturalism to theism is, eo ipso, an argument for atheism, even though such an argument need not aim to establish that naturalism is true, or rationally required, or morally required.

1. Preliminaries

Theism is the claim that there are . claims that there is just one —God; claims that there are many gods. Gods are beings or forces that have and exercise power over the natural world and that are not, in turn, under the power of higher‐ranking or more powerful categories of beings or forces. Thus, monotheism claims that there is just one supernatural being or force—God—that has and exercises power over the natural world and that is not, in turn, under the power of higher‐ranking or more powerful categories of beings or forces.

Naturalism is the claim that there are none but natural causes, beings and forces. Naturalism entails that all causally efficacious beings and forces are located within the natural world. As noted above, naturalism is inconsistent with theism: naturalism entails that there are no supernatural beings or forces that have and exercise power over the natural world, whereas theism entails that there are supernatural beings that have and exercise power over the natural world.

Supernaturalism—perhaps it might more neatly be called ‘anaturalism’—is the denial of naturalism. Just as naturalism is only one form of atheism, so, too, theism is just one kind of supernaturalism. There can be—and are—atheists who embrace the supernatural; there can be—and are— supernaturalists who do not embrace theism.

Monotheists disagree about the nature of God. Some monotheists suppose that God is personal; others do not. Some monotheists suppose that God is simple; others do not. Some monotheists suppose that God is impassible; others do not. Some monotheists suppose that God is triune; others do not. Some monotheists suppose that God is perfectly good; others do not. Etc. 2

Naturalists disagree about the nature of natural reality. Some naturalists suppose that the natural supervenes upon the microphysical; others do not. Some naturalists suppose that the natural is reducible to the physical; others do not. Some naturalists suppose that the mental is emergent relative to the biological; others do not. Some naturalists suppose that natural reality is exhausted by the spatiotemporal domain downstream from that “big bang” whose remnants can be detected by our most powerful telescopes; others do not. Etc.

When particular proponents of theism and naturalism argue with one another, they will always disagree about far more than the basic claims that are constitutive of these positions. Moreover, and consequently, when particular proponents of theism and naturalism argue with one another, the details of their arguments may have little or no wider philosophical significance. While there are interesting and important observations to be made concerning the proper conduct and regulation of these kinds of disputes, we shall instead turn out attention to the prospects of finding worthier deservers of the label “argument for atheism” in a more idealised setting.

Imagine, then, that Theist and Naturalist are parties to a philosophical debate. Theist is committed to the claim that there are gods; Naturalist is committed to the claim that there are none but natural causes. Beyond these minimal commitments, Theist and Naturalist are flexible: we can dress them up with further commitments, and see how they fare. But, whenever we do dress them up with further commitments, we should make sure that each is equipped with those further commitments to the same level of detail—and we should also make sure that we assess each view to the same theoretical standards and against the same benchmarks.

2. Direct Arguments for Atheism

One strategy that is open to Naturalist is to argue that theism fails on its own terms. In pursuing this strategy, Naturalist need not be trying to persuade Theist to adopt naturalism; the object may simply be to try to encourage Theist to give up theism.

2.1 Theism is Meaningless

At various points in the history of philosophy, there have been philosophers who have tried to argue that theism is not a meaningful hypothesis. The paradigm example is Ayer (1936). Ayer claims that the sentence ‘There exists a transcendent god’ has ‘no literal significance’ (158). In saying that this sentence—which we can take to be equivalent to the defining claim of theism, viz. that there is at least one god—has no literal significance, Ayer is saying two things: first, that this sentence is not an analytic truth—i.e. not a sentence that is true simply in virtue of the words from which it is composed—and, second, that there are no actual or possible observations that are relevant to the determination of its truth or falsehood (52).

Enthusiasm for Ayer’s position has evaporated almost entirely since the latter stages of the twentieth century. While many contributors to Mitchell (1958) and Diamond and Lizenbury (1975) essentially agreed with Ayer, it is hard to find any philosophers beyond Nielsen (1971)(1982)(1985) and Martin (1990) who endorses the claim that that the sentence ‘There is at least one god’ has no literal significance. 3

There are various good reasons for this. First, Ayer’s argument depends upon a controversial verificationist theory of meaning. While there are still some verificationist holdouts—e.g. Wright (1989)—there are many who suppose that has been decisively refuted. (Consider, for example, the argument of Lewis (1988).) Second, despite Ayer’s confident assertion, it is not entirely obvious that there are no actual or possible observations that are relevant to the determination of the truth or falsity of theism. Certainly, there are people—including trained philosophers—who claim to have had experiences that they themselves take to directly support theism. And many suppose that they can describe possible courses of experience that would provide those who underwent those courses of experiences with good reason to suppose that there is at least one god. (See, for example, Alston (1991).) Third, it is worth observing that, on Ayer’s own account, atheism and naturalism are no more literally meaningful than theism: if a sentence is meaningless, then so is the denial (negation) of that sentence, and so, too, is any sentence that entails the denial (negation) of the sentence in question. If argument for the meaninglessness of theism succeeded, it might well also establish the meaninglessness of naturalism and atheism (and hence might well not ultimately lead to a victory for Naturalist).

2.2 Theism is Incoherent

Logical positivism is not the only path that has been claimed to lead to the conclusion that there is something linguistically amiss with assertions of theistic commitment. Some philosophers of a broadly Wittgensteinian persuasion have argued that claims, affirming the existence of supernatural beings and forces that have and exercise power over the natural world, are ‘ungrammatical’ or otherwise an affront to the canons of ordinary linguistic understanding. (Many of these philosophers claim also to be friends of ; they insist that religion—properly so‐called—has no commerce with supernatural beings and forces that have and exercise power over the natural world. Since our present concern is with atheism rather than with irreligion, we need not pause to scruple.) Consider Rundle (2004:77): “I can get no grip on the idea of an agent doing something where the doing, the bringing about, is not an episode in [space and] time, something involving a[n embodied and] changing agent”.

Arguments of this kind stand or fall with their broadly Wittgensteinian philosophical underpinnings. On the one hand, they invite Russellian retort: how could profession of Wittgensteinian intellectual shortcoming be a good argument for anything at all? (“I am not responsible for your intellectual shortcomings, young man!”) On the other hand, there is a fairly widespread contemporary consensus that the broadly Wittgensteinian underpinnings cannot be satisfactorily defended: rather than suppose that most philosophy is language on holidays, contemporary philosophers are much more likely to suppose that Wittgensteinian ordinary language approaches are philosophy on holidays.

2.3 Theism is Logically Inconsistent

Many philosophers have argued that particular versions of theism are logically inconsistent. If we suppose that were God to exist, God would have a sufficiently wide range of properties—essential omniscience, essential omnipotence, essential perfect goodness, necessary existence, essential simplicity, essential impassibility, essential perfect libertarian freedom, essential consciousness, essential personality, essential foreknowledge, essential infinity, essential eternity, and so forth— 4

then there is ample opportunity to argue for the logical inconsistency of God as thus conceived. On the one hand, we might argue that, considered alone, some of the properties in question are self‐ contradictory; on the other hand, we might argue that, considered together, some subsets of the properties in question are jointly contradictory. Examples abound. Some have argued that nothing can be essentially omnipotent (e.g. Sobel (2004)). Some have argued that nothing can be essentially omniscient (e.g. Grim (1991)). Some have argued that nothing can be essentially simple (e.g. Gale (1991)). Some have argued that nothing can be essentially omnipotent and essentially perfectly good (e.g. Pike (1969)). Some have argued that nothing can have essentially perfect libertarian freedom and yet be essentially perfectly good (e.g. Rowe (2004)). Some have argued that nothing can be essentially conscious and essentially impassible (e.g. Drange (1998a)). And so forth.

There is a great deal of detailed discussion that can be given of these kinds of arguments. I shall venture just a couple of general comments here. First, it is obvious that these kinds of arguments do not target theism—i.e., they do not target the claim that there are gods. Second, many of these arguments depend upon particular analyses of the key concepts involved: particular analyses of omnipotence, omniscience, perfect goodness, simplicity, freedom, consciousness, and so forth. To the extent that these arguments do depend upon particular analyses of the key concepts involved, they are vulnerable to the response that they have simply adopted the wrong analyses of these concepts. Third, these kinds of arguments are sometimes spectacularly successful in particular local debates; and these kinds of arguments do sometimes focus on difficulties that theists have found particularly troubling. So, for example, Leibniz and Clarke disagreed about what is required to reconcile essential perfect goodness and essentially perfect libertarian freedom; the apparent conflict between essential perfect goodness and essentially perfect libertarian freedom was a genuine difficulty for them.

2.4 Theism is Impossible

Some philosophers have argued that theism is, if not logically inconsistent, at any rate, (metaphysically) impossible. If we ignore the various qualifications and hedges introduced in the text, it seems to me to be possible to read Fales (2010) as arguing for a view of this kind. Fales actually calls for a ‘re‐examination of the metaphysical and epistemological conditions that must obtain if God is to have [certain] characteristics, in the light of the best current philosophical and physical understandings of causation, laws of nature, space, time and knowledge’ (157). But the reasons that he gives for calling for this re‐examination can plausibly be marshalled to construct an argument for the claim that, given our best current philosophical and physical understanding of causation, laws of nature, space, time, and knowledge, it is simply impossible that there is an omnipotent and omniscient God. While such an argument would not target theism, it might even be possible to draw upon a subset of the considerations that he marshals in order to construct an argument for the claim that, given our best current philosophical and physical understandings of causation, laws of nature, space and time, it is simply impossible for there to be gods. (‘What a theist must reckon with, to put the matter a bit differently, is that [supernatural causation] upsets the account sheet on energy and momentum; it entails that these are not conserved.’ (154) Etc.)

I think that the best position for Naturalist to adopt is one according to which theism is impossible. All possible worlds share an initial segment with the actual world. All possible worlds evolve 5

according to the same laws as the actual world. It is impossible that the actual laws could oversee a transition from a purely natural state to a state in which there are supernatural entities. There have never been any supernatural entities. So supernatural entities are impossible; and hence, in particular, gods are impossible.

I also think that the best position for Theist to adopt is one according to which naturalism is impossible. All possible worlds share an initial segment with the actual world. All possible worlds evolve according to the same laws as the actual world. It is impossible that the actual laws could oversee a transition from a state in which there are gods to a purely natural state to a state. There have always been gods. So naturalism is impossible.

Given the symmetry of this situation, I think that the prospects of a successful argument for the impossibility of theism stand or fall with the prospects of a successful argument for the falsity of theism. (Theists deny that the natural world is a causally closed system; a fortiori, it is unsurprising that theism ‘upsets the account sheet on energy and momentum’, if that ‘account sheet’ is supposed to be exclusively naturalistic.)

2.5 Theism is Inconsistent with Known Fact

Many philosophers have argued that particular versions of theism are logically inconsistent with known fact. If we suppose that, were God to exist, God would have a particular range of properties—essential omniscience, essential omnipotence, essential perfect goodness, necessary existence, essential simplicity, essential impassibility, essential perfect libertarian freedom, essential consciousness, essential personality, essential foreknowledge, essential infinity, essential eternity, and so forth—then there is ample room to argue that God’s existence is logically inconsistent with facts about the world that are acknowledged on (almost) all sides—that there is evil, that there is moral evil, that there is a lot of evil, that it is not obvious that God exists, that there are many people who fail to believe that God exists, and so forth. Some have argued that, if God existed, God would have made a world in which everyone always freely chooses the good (e.g. Mackie (1955)). Some have argued that, if God existed, God would have made God’s existence (more) obvious to all (e.g. Schellenberg (1993)). Some have argued that, if God existed, God would have ensured that all human beings came to believe in God before they died (e.g. Drange (1998b)). And so forth.

These arguments do not target theism. Indeed, most of these arguments target only a particular version of monotheism: one on which it is supposed that God is, at least, omnipotent, omniscient, and perfectly good. This is not to deny that these arguments have local significance: there are, after all, many theists who claim to believe that God is omnipotent, omniscient, and perfectly good. Moreover, there are many who also claim that, if God were not omnipotent, omniscient and perfectly good, then God would not be worshipworthy, i.e. not an appropriate focus for religious veneration. However, the point remains that these arguments have far more significance for a particular brand of theists than they do for any atheists: even if arguments of this kind are successful, they certainly do not succeed in showing that there are no gods. And, of course, it is controversial whether these kinds of arguments do succeed (but, of course, a detailed examination of these arguments is beyond the scope of the present chapter).

2.6 Theism is Improbable Given Known Fact 6

Many philosophers have argued that particular versions of theism are improbable in the light of known facts. If we suppose that, were God to exist, God would have a particular range of properties—essential omniscience, essential omnipotence, essential perfect goodness, necessary existence, essential simplicity, essential impassibility, essential perfect libertarian freedom, essential consciousness, essential personality, essential foreknowledge, essential infinity, essential eternity, and so forth—then there is ample room to argue that God’s existence is improbable in the light of facts about the world that are acknowledged on (almost) all sides—that there are horrendous evils, that there are evils for which we are unable to identify outweighing goods, that the universe does not appear to have a ‘human scale’, and so forth. Some have argued that it is improbable that, if God existed, God would have permitted certain kinds of horrendous evils (e.g. Rowe (1979)). Some have argued that it is improbable that, if God existed, God would have created a universe in which the domain of humanity is so insignificant (e.g. Everitt (2004)). Some have argued that it is improbable that, if God existed, God would have produced such biologically suboptimal creatures as human beings (e.g. Dawkins (1986)). Some have argued that it is improbable that, if God existed, God would have created a world in which there is the distribution of pain and pleasure in sentient creatures that we find in the actual world (e.g. Draper (1989)). And so forth.

These arguments do not target theism. Indeed, most of these arguments target only a particular version of monotheism: one on which it is supposed that God is, at least, omnipotent, omniscient, and perfectly good. This is not to deny that these arguments have local significance: there are, after all, many theists who claim to believe that God is omnipotent, omniscient, and perfectly good. Moreover, there are many who also claim that, if God were not omnipotent, omniscient and perfectly good, then God would not be worshipworthy, i.e. not an appropriate focus for religious veneration. However, the point remains that these arguments have far more significance for a particular brand of theists than they do for any atheists: even if arguments of this kind are successful, they certainly do not succeed in showing that it is improbable that there are gods. And it is controversial whether any of these kinds of arguments succeed (either separately or jointly).

2.7 Theism is Morally Repugnant

While many people who have rejected theism have regretted their inability to believe in God or the gods, there are some people who have supposed that the theoretical unacceptability of theism dovetails nicely with its moral unacceptability. If we suppose that, were God to exist, God would have a particular range of properties—selected from among essential omniscience, essential omnipotence, essential perfect goodness, necessary existence, essential simplicity, essential impassibility, essential perfect libertarian freedom, essential consciousness, essential personality, essential foreknowledge, essential infinity, essential eternity, and so forth—then there is at least some room to argue that God’s existence is morally undesirable and perhaps even morally repugnant. For instance, one might argue as follows: The only kind of freedom that it is possible to have is compatibilist freedom. But it is impossible to have compatibilist freedom if there is a causally upstream agent who selects one’s beliefs and desires. So it is impossible for you to be free if you are one of God’s creatures. But freedom is a highly significant moral good. So God’s non‐existence is morally desirable: God’s non‐existence is necessary for our freedom and the goods that our freedom makes possible—e.g. moral responsibility. (See Kahane (2011) for other arguments along similar lines.) 7

As in the previous two cases, these arguments do not target theism. Thus, for example, the sample argument that I have given only targets versions of monotheism that suppose that there is a strong sense in which God creates us. Moreover, even in the context of debate with theists who do make the relevant assumptions, it is not clear how much weight these kinds of arguments could carry: after all, even if it were true that theism is morally repugnant, that, in itself, would certainly not be a good reason to suppose that it is false.

2.8 Finishing Touches

This survey of direct arguments for atheism has been very brief, and has certainly not mentioned— let alone considered—the wide range of direct arguments for atheism that are to be found in the literature. While I am not, myself, particularly enthusiastic about the prospects for successful direct arguments for atheism, I think that it is clear that there is a great deal more to be done in clarifying and analysing the arguments that can be put forward in the various categories that I have identified (and perhaps also in categories to which I have not attended), and also in thinking about the ways in which some of these arguments might be combined to form direct ‘cumulative’ arguments for atheism.

In closing, there is perhaps one more gambit that deserves some mention. Some people suppose that there is a standing presumption against existence claims. So, for example, such people might suppose that, prior to examination of the evidence, there is a standing presumption that there is no china teapot in orbit around Pluto. But, if that’s right, then such people might further suppose that all that one needs in order to produce a good argument for atheism is to produce good objections to all of the arguments that can be offered for theism. If no argument for theism succeeds—as argued in, for example, Oppy (2006)—then the standing presumption against existence claims kicks in, and one has good reason to accept atheism. For myself, I do not think that there is a standing presumption against existence claims; I do not think that considerations about burden of proof have a significant role in the arbitration of our dispute between Naturalist and Theist. (For more about the proper conception of the dispute between Naturalist and Theist, see Oppy (2011).)

3. Comparative Arguments for Atheism

A different strategy that is open to Naturalist is to argue that naturalism is theoretically superior to theism. The idea here is to compare the theoretical merits of naturalism with the theoretical merits of theism when these views are assessed against the relevant available evidence.

In order to proceed, we need to have some conception of theoretical merit. While this matter remains controversial, there is fairly broad consensus that appropriate trade off of complexity of theory with fit with data, breadth of explanatory role, and compatibility with independently established theory are amongst the considerations that are to be weighed in any assessment of the merits of competing theories. While there are competing views about how to measure the complexity of a theory, I shall suppose that relevant factors include: numbers and kinds of primitive terms; numbers and kinds of primitive predicates; and numbers and kinds of other theoretical primitives (e.g. sentential operators). 8

We proceed to consider how theism and naturalism measure up against these theoretical desiderata given various key pieces of evidence.

3.1 Ultimate Explanation

We take as our first piece of evidence the existence of a global (efficient) causal order. Given that there is a global causal order, we can frame various hypotheses about its shape: (1) infinite regress; (2) necessary initial state; (3) contingent initial state (involving some necessary existents); (4) contingent initial state (involving only contingent existents). We can then assess the theoretical credentials of naturalism and theism against these various hypotheses.

Naturalism says that there is only the natural causal order: the ordered global causal states of the natural world. Theism says that there is the natural causal order, and more besides: there is the supernatural causal order, and there is causal commerce between the natural and the supernatural. Taking only considerations of theoretical simplicity into account, it is clear that naturalism is ahead: it postulates fewer kinds of entities, fewer kinds of causes, and so forth. Moreover, when we turn to consider questions of ultimate explanation—Why is there something rather than nothing? Why is there a causal order? Why is there a natural causal order?—it is clear that theism gains no advantage over naturalism. For, whatever answer to these questions turns out to be correct—Because there always has been! [Infinite Regress]; Because there had to be this particular initial causal state! [Necessary Initial State]; Because there had to be some initial causal state or other! [Contingent Initial State Involving Some Necessary Existents]; Just because! [Contingent Initial State Involving Only Contingent Existents]—naturalism supports that answer at least as well as theism does. So, given that we consider only questions about ultimate explanation, naturalism trumps theism.

Might there be explanations of the existence of the global (efficient) causal order that I haven’t considered? I think it unlikely. Some suppose that the existence of the global (efficient) causal order might have an axiarchial explanation: there is a global (efficient) causal order because it is good that there be such a global (efficient) causal order (see, for example, Leslie (1979)). However, I’m happy to rule this attempt out of court: it is impossible for the existence of the global (efficient) causal order to be explained in this way. And—at least at the time of writing—there are no other contending explanations for the existence of the global (efficient) causal order that have come into view.

3.2 Order

We take as our second piece of evidence the alleged fine‐tuning for life of the domain in which we live. Although it is controversial whether the domain in which we live is fine‐tuned for life, we shall, for the sake of argument, simply suppose that the domain in which we live is fine‐tuned for life.

There are two hypotheses that we can frame about the point in the causal order at which the fine‐ tuning for life of the domain in which we live was fixed: either it has been fixed at all points in the causal order that the domain which we live is fine‐tuned for life; or else there is some initial segment of the causal order in which it is not fixed that the domain in which we live is fine‐tuned for life. 9

We can now assess the theoretical credentials of naturalism and theism against the sum of these hypotheses concerning the point in the causal order at which the fine‐tuning for life of the domain in which we live was fixed and the previous hypotheses about the shape of the global causal order.

As before, it is clear that, taking only considerations of theoretical simplicity into account, naturalism is ahead. We have already seen that, when we turn to consider questions of ultimate explanation, theism gains no advantage over naturalism. But it is equally clear that adding questions about the point in the causal order at which the fine‐tuning for life of the domain in which we live was fixed also creates no advantage for theism over naturalism. On the one hand, if it has been fixed at all points in the causal order that the domain in which we live is fine‐tuned for life, then there is just the same range of explanatory options available to naturalism as there are available to theism: Because the causal order has always been fine‐tuned for life! [Infinite Regress]; Because there had to be this particular initial causal state and it had to be fine‐tuned for life! [Necessary Initial State]; Because there had to be some particular initial causal state that had to be fine‐tuned for life! [Contingent Initial State Involving Essential Fine‐Tuning]; Just because! [Contingent Initial State Involving Inessential Fine‐Tuning]. On the other hand, if there is some initial segment of the causal order in which it is not fixed that the domain in which we live is fine‐tuned for life then, again, there is just the same range of explanatory options for naturalism as there are for theism: for, in this case, it can only be that it is a matter of objective chance that the domain in which we live is fine‐tuned for life. So, given that we consider only questions about ultimate explanation and fine‐tuning (ultimate order), naturalism trumps theism.

3.3 The Necessary and Knowable A Priori

We take as our third piece of evidence the various domains that have often been taken to be home to claims and/or entities that are necessary and knowable a priori: at least logic and mathematics; and perhaps also some or all of modality, morality, meaning, (analytic) metaphysics, and so forth. Whatever among these domains is home to claims and/or entities that are necessary and knowable a priori is such that it is home to claims that are true at all points in the causal order and/or to entities that exist at all points in the causal order. So, no matter what hypothesis we make about the shape of the global causal order, whatever among these domains is home to claims and/or entities that are necessary and knowable a priori is home to claims and/or entities that are explained just as well on naturalism as they are on theism. After all, whatever is true at all points in the causal order and/or whatever exists as all points in the causal order has no cause: thus, whatever is true at all points in the causal order and/or exists at all points in the causal order is theoretically primitive, at least relative to all domains that are not home to claims and/or entities that are necessary and knowable a priori—though, of course, for example, some mathematical claims and/or entities may be explained in terms of other mathematical claims and/or entities, or in terms of logical claims and/or entities, or in terms of the claims and/or entities of some other domain that is necessary and knowable a priori.

It will of course be noted that the argument just given relies on the assumption that theism is not necessary and knowable a priori: theism does not have the same standing as whichever among logic, mathematics, modality, morality, meaning, (analytic) metaphysics, and so forth, has that standing. 10

Some may dispute this assumption (even though I take it to be obviously correct); however, further argument over this point will need to be deferred to some other occasion.

Finally, of course, it should also be noted that it is controversial whether any of the domains mentioned in the above argument is necessary and knowable a priori. The aim of the argument is only to establish that, insofar as there are domains that are home to claims and/or entities that are necessary and knowable a priori, those domains offer no support to theism over naturalism. Given that we consider only questions about ultimate explanation, fine‐tuning (ultimate order), and the necessary and knowable a priori, naturalism trumps theism.

3.4 (Objective) Value

We take as our fourth piece of evidence the various axiological domains: moral, aesthetic, comic, and so forth. There are various hypotheses that we can frame about the way in which values are related to the causal order; however, the ones that are most hospitable to theism—and the ones that we shall consider henceforth—suppose that values have some kind of objective connection to the causal order: causal agents, states, events and processes have objective values (even though, of course, objective values are not themselves causal agents, states, events or processes).

If we suppose that values have some kind of objective connection to the causal order, then it seems to me to be inevitable that we also suppose that the connection in question is necessary. For naturalists, this is surely obvious: how could it be that purely naturalistic causal agents, states, events and processes have objective values unless there is some kind of necessary connection between those values and the ways that those purely naturalistic causal agents, states, events, and processes are? But it seems no less evident for theists. If supernatural causal agents, states, events and processes have objective values, then there is just the same reason to suppose that there is some kind of necessary connection between those values and the ways that those supernatural causal agents, states, events and processes are. But once we have accepted that there are (primitive) necessary connections between objective values and some causal agents, states, events and processes, there is only theoretical loss in failing to suppose that those same (primitive) necessary connections hold for objective values and all other causal agents, states, events and processes. Put another way: either theism and naturalism are on a par—because each accepts the same kind of relationship between objective values and causal agents, states, events and processes—or else naturalism should be preferred to theism because it gives a simpler (unified) account of the relationship between objective values and causal agents, states, events and processes.

The argument just given is obviously related to the argument attributed to Socrates by Plato at Euthyphro 10a‐11b. If some part of the causal order is supposed to establish objective connections between values and the causal order, then there is an obvious difficulty that arises concerning the objectivity of the values that apply to that very part of the causal order. No part of the causal order can come to possess an objective value merely by deeming itself to be in possession of that value; in particular, no supernatural agent or power can come to possess an objective value merely by deeming itself to be in possession of that value.

Of course, it is controversial to suppose that there are objective values. However, if there are no objective values, then there is surely no reason at all to suppose that considerations about values 11

might somehow favour theism over naturalism. Given that we consider only questions about ultimate explanation, fine‐tuning (ultimate order), the necessary and knowable a priori, and (objective) value, naturalism trumps theism.

3.5 Meaning

We take as our fifth piece of evidence the alleged meaninglessness of the causal order in the absence of the supernatural. Many people have claimed that, in the absence of certain kinds of supernatural agents—God, or the gods—the causal order would have no meaning or purpose, that this would be a tragic and regrettable state of affairs, that in the presence of those same kinds of supernatural agents—God, or the gods—the causal order does have meaning and purpose, and that this is a state of affairs that should be celebrated and applauded.

This can’t be right. If the thoughts, feelings, and deeds of human beings are not sufficient to endow their part of the natural order with meaning, then how could the thoughts, feelings, and deeds of supernatural powers and agents be sufficient to endow a wider causal order with meaning? If a causal order can only be imbued with meaning from an external source—something outside the causal order in question—then, of course, it will be true that the natural causal order is devoid of meaning; but it will be equally true that, if a causal order can only be imbued with meaning from an external source—something outside the causal order in question—then any larger causal order within which the natural causal order is embedded will also be devoid of meaning.

Perhaps we can also add this: Suppose that it were true that, if there were only the natural causal order, then the natural causal order could have no meaning or purpose. How, then, could it be true that adding some further part to the causal order—where that further part of the causal order itself had no meaning or purpose—somehow gave meaning or purpose to the natural causal order? Perhaps it might be suggested that there could be a kind of causal order that gave meaning and purpose to itself: a kind of intrinsically meaningful causal order. But if we suppose that supernatural powers and agents have the capacity to endow the causal order to which they belong with meaning and purpose, surely we can also suppose that natural agents have the capacity to endow the causal order to which they belong with meaning and purpose.

The upshot here is clear: considerations about the meaningfulness or otherwise of the causal order confer no advantage on theism over naturalism. Given that we consider only questions about ultimate explanation, fine‐tuning (ultimate order), the necessary and knowable a priori, (objective) value, and the meaningfulness of the causal order, naturalism trumps theism.

3.6 Consciousness

We take as our sixth piece of evidence the presence in the natural universe of entities—human beings—that have conscious mental lives. Some people have claimed that theism gives a better explanation than naturalism does of the presence in the natural universe of entities with conscious mental lives. (See, for example, Moreland (2008).) While I don’t really have space here to do full justice to the relevant considerations, I can at least sketch an argument for the rejection of this claim.

Naturalism is committed to affirming some kind of relation or connection between conscious states and natural states (in human beings, neural states). Opinion varies concerning the nature of this 12

relation or connection: it might be identity, or some kind of necessity, or some kind of supervenience that falls short of necessity, or some kind of emergence that falls short of supervenience. But what options does theism encompass? If theism supposes that there is no relation or connection between conscious states and supernatural states, then it plainly has no explanatory advantage over naturalism. But if theism says that there is a connection between conscious states and supernatural states (say, states of God, or gods, or supernatural ), then it seems to have the same range of available options concerning the nature of this relation or connection in the domain of the supernatural: it might be identity, or some kind of necessity, or some kind of supervenience that falls short of necessity, or some kind of emergence that falls short of supervenience. Since there is plainly no explanatory advantage that accrues to theism, considerations of simplicity favour naturalism, no matter how we suppose that the chips fall concerning the relationship or connection between conscious states and states in the relevant causal domain.

Perhaps it is worth noting here that on my favourite theory of the range of possibility (sketched in Section 2.4 above), zombies are impossible: there are no possible worlds in which the kinds of neural states that are actually—or are actually correlated with—conscious states fail to be—or to be correlated with—conscious states. My preferred position is that conscious states in human beings just are neural states; I recommend this view to naturalists, but do not suppose that they are obliged to adopt it. In any case, I conclude that, given that we consider only questions about ultimate explanation, fine‐tuning (ultimate order), the necessary and knowable a priori, (objective) value, the meaningfulness of the causal order, and consciousness, naturalism trumps theism.

3.7 Reason

We take as our seventh piece of evidence the presence in the natural universe of entities—human beings—that have higher‐order cognitive capacities: the abilities to reason, infer, argue, and so forth. Some people have claimed that theism gives a better explanation than does naturalism of the presence in the natural world of beings that have higher‐order cognitive capacities. (See, for example, Lewis (1947) and Reppert (2003).) Once more, while I don’t really have space here to do full justice to the relevant considerations, I can at least sketch an argument for the rejection of this claim.

Naturalism is committed to affirming some kind of relation or connection between episodes of reasoning, inferring, arguing, etc. and natural episodes (in human being, episodes intimately involving neural states). Opinion varies concerning the nature of this relation or connection: it might be identity, or some kind of necessity, or some kind of supervenience that falls short of necessity, or some kind of emergence that falls short of supervenience. But what options does theism encompass? If theism supposes that there is no relation or connection between episodes of reasoning, inferring, arguing, etc. and supernatural episodes (episodes involving God, or gods, or supernatural souls), then it plainly has no explanatory advantage over naturalism. But if theism says that there is a connection between episodes of reasoning, inferring, arguing, etc. and supernatural episodes (say, states of God, or gods, or supernatural souls), then it seems to have the same range of available options concerning the nature of this relation or connection in the domain of the supernatural: it might be identity, or some kind of necessity, or some kind of supervenience that falls short of 13

necessity, or some kind of emergence that falls short of supervenience. Since there is no explanatory advantage that accrues to theism, considerations of simplicity favour naturalism, no matter how we suppose that the chips fall concerning the relationship or connection between episodes of reasoning, inferring, arguing, etc. and episodes in the relevant causal domain.

I conclude that, given that we consider only questions about ultimate explanation, fine‐tuning (ultimate order), the necessary and knowable a priori, (objective) value, the meaningfulness of the causal order, consciousness, and reason, naturalism trumps theism.

3.8 Supernatural Experience

We take as our eighth piece of evidence to be the directly and indirectly recorded supernatural experience of humanity, from its historical origins—whatever those might have been—to the present. Many people across history have claimed to have direct experiences of supernatural agents: experiences that the people in question took to be of, or to be directly caused by, supernatural agents—God, gods, angels, demons, ghosts, ancestor spirits, fairies, and so forth.

Without pretending to be able to make a full and proper assessment of the evidential weight of supernatural experience here, we can note a few salient points. First, there is no question that the history of reports of encounters with supernatural beings and forces is, at least in very large part, a history of fraud, gullibility, deception, stupidity, ignorance, and so forth. Second, there is no serious doubt that there is at least good prima facie reason to believe that there is a huge panoply of supernatural beings whose existence would be vindicated by the recorded supernatural experience of humanity if the existence of any supernatural beings was vindicated by that recorded supernatural experience. Third, it is quite clear that the joint effect of these first two points is to raise serious questions about the evidential worth of any reports of experiences that are claimed to be of, or directly caused by, supernatural agents. Fourth, it may well be that, in the absence of defeating considerations, its seeming to someone to be the case that p is some reason to suppose that it is the case that p (cf. Swinburne (1979)). But, as we have just noted, there is no serious doubt that there are very weighty candidate defeating considerations in the case of ‘seemings’ that are tied to the supernatural.

In the absence of any independent support for in gods—i.e. support founded in something other than reports of experiences that have been taken to be of, or directly caused by, gods—there is clearly reason to prefer the uniform treatment of reports of supernatural experiences that naturalism affords to the non‐uniform treatment of reports of supernatural experiences that is required by any developed version of theism. That is: given that we consider only questions about ultimate explanation, fine‐tuning (ultimate order), the necessary and knowable a priori, (objective) value, the meaningfulness of the causal order, consciousness, reason, and supernatural experience, naturalism trumps theism.

3.9 (Supernatural) History

We take as our ninth piece of evidence the history of our universe, as recorded in our best scientific and historical theories. This history contains numerous direct and indirect reports of—and speculations about—supernatural engagement with the natural world. Some people suppose that 14

the biological sciences furnish us with good reasons to suppose that a supernatural intelligent designer interfered at various points in the evolution of life in our universe. Some people suppose that various reported belonging to particular religious traditions furnish us with good reasons to suppose that those reported miracles actually involved supernatural agency. Some people suppose that the contents of the scriptures of particular religious traditions furnish us with good reasons to suppose that those scriptures are actually direct products of supernatural agency. And so forth.

Again, we can do no more than note a few salient points here. First, what goes for reports of supernatural experience goes equally for ancient written reports—and, indeed, reports more generally—of the deeds of supernatural agents: there is no question that, at least in very large part, we are in the domain of fraud, gullibility, deception, stupidity, ignorance, and so forth; and there is also at least good prima facie reason to think that there is a huge panoply of supernatural beings whose existence would be vindicated by reports of the deeds of such agents if the existence of any supernatural beings was vindicated by those kinds of reports. Second, the track record of the supernatural in science is null and void: there is not one established scientific result that depends upon the postulation of the existence of something supernatural, and the history of attempts to repair alleged ‘gaps’ in natural science with supernatural filler is a long record of dismal failure.

In the absence of any independent support for belief in gods—i.e. support founded in something other than (a) reports of experiences that have been taken to be of, or directly caused by, gods; (b) reports—ancient or recent, written or verbal—of interventions by gods in the natural order; and (c) attempts to improve upon natural science by appeal to the activities of gods—there is clearly reason to prefer the uniform treatment of these topics that is afforded by naturalism to the non‐uniform treatment of these topics that is provided by any developed version of theism. That is: given that we consider only questions about ultimate explanation, fine‐tuning (ultimate order), the necessary and knowable a priori, (objective) value, the meaningfulness of the causal order, consciousness, reason, supernatural experience, and (supernatural) history, naturalism trumps theism.

3.10 Finishing Touches

Clearly, the preceding sketch of a comparative argument for atheism is incomplete. First, the treatment of each of the relevant pieces of evidence is incomplete. Second, not all of the relevant evidence has been considered. And, third, there is clearly more work to be done on the theory of theoretical virtue that underpins the argument. While I can imagine that there is a way of filling out this sketch that justifies the conclusion that, given all of the available evidence, naturalism trumps theism, I am sceptical that the task can be completed (even in principle). However, I do think that this kind of comparative argument marks the most promising current strategy for arguing for atheism: if there is a successful argument for atheism, then, I think, it will run along something like these lines .

4. Concluding Remarks

There are many loose ends in this discussion. I shall tie just one here. Given the way that I set up the initial discussion, it turns out—I think—that is a version of naturalism. Moreover, it turns out that there are self‐identified ‘theists’ who are classified as ‘naturalists’ in the above discussion— 15

though, interestingly, these ‘theists’ also self‐identify as ‘naturalists’: see, for example, Forrest (1996) and Bishop (2007). This means that—setting disputes about the proper meanings of terms to one side—there is still a further argument to be had, within what I have called the ‘naturalist’ camp— between those who suppose that the natural world has the kinds of properties that those naturalists who call themselves ‘theists’ or ‘pantheists’ attribute to it, and those who suppose that the natural world does not have the kinds of properties that those naturalists who call themselves ‘theists’ or ‘pantheists’ attribute to it. In that debate, I certainly come done on the latter side—see, for example, Oppy (1997)—but there is no space to rehearse the relevant considerations here.

References

Alston, W. (1991) Perceiving God: The Epistemology of Religious Experience Ithaca: Cornell University Press

Ayer, A. (1936) Language, Truth and Logic London: Victor Gollancz

Bishop, J. (2007) Believing by : An Essay in the Epistemology and Ethics of Religious Belief Oxford: Oxford University Press

Dawes, G. (2009) Theism and Explanation New York: Routledge

Dawkins, R. (1986) The Blind Watchmaker Harlow: Longman

Diamond, M. and Lizenbury, T. (eds.) (1975) The Logic of God Indianapolis: Bobbs‐Merrill

Drange, T. (1998a) ‘Incompatible‐Properties Arguments: A Survey’ Philo 2, 49‐60

Drange, T. (1998b) Non‐Belief and Evil Amherst: Prometheus Books

Draper, P. (1989) ‘Pain and Pleasure: An Evidential Problem for Theists’ Noûs 23, 331‐50

Everitt, N. (2004) The Non‐ London: Routledge

Fales, E. (2010) Divine Intervention: Metaphysical and Epistemological Puzzles New York: Routledge

Forrest, P. (1996) God without the Supernatural Ithaca: Cornell University Press

Gale, R. (1991) On the Nature and Existence of God Cambridge: Cambridge University Press

Grim, P. (1991) The Incomplete Universe: Totality, Knowledge and Truth Cambridge: MIT Press

Kahane, G. (2011) ‘Should we want God to Exist?’ Philosophy and Phenomenological Research 82, 674‐96

Leslie, J. (1979) Value and Existence Oxford: Blackwell

Lewis, C. (1947) Miracles: A Preliminary Study New York: Macmillan

Lewis, D. (1988) ‘Statements Partly About Observation’ Philosophical Papers 17, 1‐31

Mackie, J. (1955) ‘God and Omnipotence’ Mind 64, 200‐12 16

Martin, M. (1990) Atheism: A Philosophical Justification Philadelphia: Temple University Press

Mitchell, B. (1958) Faith and Logic London: Allen & Unwin

Moreland, J. (2008) Consciousness and the Existence of God: A Theistic Argument New York: Routledge

Nielsen, K. (1971) Contemporary Critiques of Religion New York: Herder & Herder

Nielsen, K. (1982) An Introduction to New York: St. Martin’s Press

Nielsen, K. (1985) Philosophy and Atheism Buffalo: Prometheus Books

Oppy, G. (1996) Ontological Arguments and Belief in God Cambridge: Cambridge University Press

Oppy, G. (1997) ‘Pantheism, Quantification, and Mereology’ Monist 80, 320‐36

Oppy, G. (2006) Arguing about Gods Cambridge: Cambridge University Press

Oppy, G. (2011) ‘Über die Aussichten erfolgreicher Beweise für Theismus oder Atheismus’ in J. Bromand and G. Kreis (eds.) Gottesbeweise von Anselm bis Gödel Berlin: Suhrkamp Verlag

Pike, N. (1969) ‘Omnipotence and God’s Ability to ’ American Philosophical Quarterly 6, 208‐12

Reppert, V. (2003) C. S. Lewis’ Dangerous Idea: In Defense of the Argument from Reason Downer’s Grove: Inter‐Varsity Press.

Rowe, W. (1979) ‘The and Some Varieties of Atheism’ American Philosophical Quarterly 16, 335‐41

Rowe, W. (2004) Can God be Free? Oxford: Oxford University Press

Rundle, B. (2004) Why There is Something rather than Nothing Oxford: Oxford University Press

Schellenberg, J. (1993) Divine Hiddenness and Human Reason Ithaca: Cornell University Press

Sobel, J. (2004) Logic and Theism: Arguments for and against Beliefs in God Cambridge: Cambridge University Press

Swinburne, R. (1979) The Existence of God Oxford: Clarendon

Wright, C. (1989) ‘The Verification Principle: Another Puncture—Another Patch’ Mind 98, 611‐22